K_multiprocessor Group (KMG) optimal scheduling algorithm for reducing preemptions and migrations in RTS

KMG is a proposed multiprocessor scheduling algorithm which is optimal for the scheduling of real time periodic tasks with implicit deadlines. It is adhering to the DPF (Deadline Partitioning Fair) principle. It was open in latest studies that the division of the time in slices bounded by two succeeding deadlines and the systematic execution of migratory tasks in each time slice innate in DPF algorithms reduce the practicality of EKG (EDF with task splitting and k processors in a Group.). But, its semi-partitioned come close to allows to bound the amount of migrating tasks and increases the area of the tasks memory, then reducing time overheads constrained by task preemptions and migrations. Therefore, we propose two techniques with the aim of lowering the amount of preemptions and migrations incite by the system when scheduled with EKG, though maintaining the benefit of its approach. The first developments that exchange execution time between tasks and time slices named as swapping algorithm. The next one is to reducing the number of time slices required to make sure that all job deadlines are valued. Both have a strong result on number of preemptions and migrations as maintaining the optimality of EKG.